* it r rw eop 5 4op
AD-A203 891
IDA MEMORANDUM REPORT M-548
CENTRAL RESEARCH PROJECT REPORT ONSUPERCONDUCTIVITY (1988):
PART I, MATERIALS AND IR DETECTORS
William S. HongJohn E. Hove
Mark S. Taylor
December 1988
D
FEB0 79m,
S INSTITUTE FOR DEFENSE ANALYSES180i N. Beauregard Street, Alexandria, Virginia 22311
89 2 ,o HOQ 3 O
DEFINMTONSIDA publishes Mhe following deomoets to repoth Uns eult of It week.
ReportsRepot ar do ma ootauheltative and most curelully cons~ere prouct IDA puhilahe.maoy nermalleoy ~ ssult of major projects which ia) hay a diret hearing on deonlo
adctn Msl programs, of (b) sddre. ssue m sfWIfan csrR to tOneacullnlandc, lOs Congres andor Mhe poblc, or (t) addres Imsm tha have * alg Mai economicImplication. IDA Reports are reviwed by outide pnes of expelt to esure thei highqWal sod ref leeftie prais 00udid, ad tihyar released by Me5 Presdt fIDA.
PapersPapers normally addrs relatively rstrced technial or policy Issues. They commniacatetoe reslt of speidl anales, Intrm report or phase of a task, ad hoc or quick reactonmeek papers are rviewsd to amn e thtey mest standad Similar to hose eIpcte ofreere IPapers In professional Journals.
Memorrundum Report.IDA Mommoandum Report am sd for toe cenvelence of dothe sp asr or te analysts toreseed ' -tas I work dean In quic recto studies OW MWo Initocte ecH nl sullppedtsctiwlt to mco avellab prellmlny and tenttiv reult of salyse or of -wkngro"p aSo Pod aclVlje to leward lnlrmalls tatcasslll nlzdsd isnesci-t,set*d or in wm a record of contorences. meetngs, or br~ingsp or of da tevelope Inthe cou of en luestlioe. ReWow of Needm Report Is sud to tecntnad Intemided am.
ma resuts of IDA work are alss conveyfed by br~ingso and iaomal memoranda to spnsrMWd "Nor denated by toe specer, when appropriate.
mae work reporte In ths oumn was coeducto mader IDA's edpedc Roeach Pro-gram. t publiale dons no Imply endorsement by toe Dapeehu of Detemes or soy othergo~e'nen qg.oy nor should the contents he constue a refiesing tOe elfclal posetived" of anygv onetaecy.
Thmis deoumenst Is -uissile and sutbefor publc release.
UNCLASSIFIEDSECURITY CLASSIFICATION OF THIS PAGE
REPORT DOCUMENTATION PAGEla. REPORT SECURITY CLASSIFICATION lb. RESTRICTIVE MARKINGS
UNCLASSIFIED
2a. SECURITY CLASSIFICATION AUTHORITY 3. DISTRIBUTION/AVAILABILITY OF REPORTN/A This document is unclassified and suitable for
2b. DECLASSIFICATION/DOWNGRADING SCHEDULE public release.N/A
4. PERFORMING ORGANIZATION REPORT NUMBER(S) . MONITORING ORGANIZATION REPORT NUMBER(S)
* IDA Memorandum Report M-5486a. NAME OF PERFORMING ORGANIZATION 6b. OFFICE SYMBOL 7a. NAME OF MONITORING ORGANIZATIONInstitute for Defense Analyses (It applicable)
6c. ADDRESS (City, State, end Zip Code) 7b. ADDRESS (CITY, STATE, AND ZIP CODE)1801 N. Beauregard StreetAlexandria, VA 22311-1772
a. NAME OF FUNDING/SPONSORING bOFFICE SYMBOL 9. PROCUREMENT IN'RUMENT IDENTIFICATION NUMBERIDA Independent Research Program
Sc. ADDRESS (City, Stae, and Zip Code) 10. SOURCE OF FUNDING NUMBERS
PROGRAM PROJECT TASK NO. WORK UNITELEMENT NO. ACCESSION NO.
CRP
11. TITLE (Include Security Claeiflcation)
Central Research Project Report on Superconductivity (1988): Part I, Materials and IR Detectors12. PERSONAL AUTHOR(S).
William S. Hong, John E. Hove, Mark S. Taylor13. TYPE OF REPORT 3b. TIME COVERED 14. DATE OF REPORT (Yew, Month, Day) 15. PAGE COUNT
Final FROM 3/88 TO 9/88 December 1988 3716. SUPPLEMENTARY NOTATION
17. COSATI CODES 16. SUBJECT TERMS (Continue an reverse If neceary and Identify by block number)
FIELD GROUP SUB-GROUP Superconductivity, Ceramic superconductors, Infrared detectors,Bolometers, Superconductor IR detectors, Phase Equilibria,
- Ceramic crystal structures, Superconducting electronic circuits
19. ABSTRACT (Continue an revere. If necessary and Identify by block number)
This report reviews and summarizes available information on the phases and crystal structures of known andpredicted high-temperature superconductors (HTS). Materials involving thallium have already exhibited criticaltemperatures up to 125 K, and techniques have been suggested to increase this value. Recent theoretical efforts
• predict that room-temperature superconducting materials, if possible at all, may require different chemicalcompositions than those oxides presently under investigation. A second part of the report reviews te possible directuse of HTS as infrared sensor elements. This indicates little or no advantage over semiconductors; however, HTSuse in the signal-processing circuitry of detector arrays may be highly beneficial. A second (later) report will describea quantitative analysis on the feasibility of a self-guided probe weapon using superconducting magnetic fieldgradient sensors. While additional analysis and calculations are still necessary, the tentative conclusion of this Part IIreport is that an HTS seeker and guidance system may be feasible for the destruction of tanks, ships, submarines, or
* other large steel targets. Such a missile would be generally insensitive to the environment or commoncountermeasures such as flares, camouflage, or smoke screens.
20. DISTRIBUTION/AVAILABILITY OF ABSTRACT J0 UNCLASSIFIED/UNLIMITED 00 SAME AS RPT. 03 OTIC USERS UNCLASSIFIED
* 22a. NAME OF RESPONSIBLE INDIVIDUAL 22b. TELEPHONE (Include Area Code) 22c. OFFICE SYMBOL
John E. Hove (703) 578-2869
DO FORM 1473. 84 MAR 83 APR edition may be ued until exhausted.All other edition* are obsolete SECURITY CLASSIFICATION OF THIS PAGE
UNCLASSIFIED
IDA MEMORANDUM REPORT M-548
CENTRAL RESEARCH PROJECT REPORT ONSUPERCONDUCTIVITY (1988):
* PART I, MATERIALS AND IR DETECTORS
William S. HongJohn E. Hove
Mark S. Taylor
S
December 1988
S
0
IDAINSTITUTE FOR DEFENSE ANALYSES
IDA Independent Research Program
PREFACE
* This Memorandum Report presents partial results of Central Research Project
(CRP) studies in the area of high-temperature superconductivity performed during FY
1988. An earlier exploratory study (Ref. 1) provided a brief background on the
characteristics of superconductors and some examples of their possible defense
* applications.
For the FY 1988 CRP effort, two activities are reported herein:
(1) Phase equilibria and crystal structure of known high-temperature super-conductors, and
(2) Applications to infrared sensor systems.
In a later report an analysis of gradiometer probe use in missile guidance systems for anti-
tank and anti-ship applications will be described.
* Some portions of the work were partially supported under two tasks from OSD in
FY 1988 (T-D2-509, Army Sensor Applications, and T-D2-566, Superconductivity
Issues). The results will be reported to those sponsors with, of course, appropriate credit
to the CRP support.
0 This report has not been subjected to formal review.
, C r,3 1 tC; t01
T SIS ' "000r
.113• --
*.
ii
O
ABSTRACT
* This report reviews and summarizes available information on the phases and
crystal structures of known and predicted high-temperature superconductors (HTS).
Materials involving thallium have already exhibited critical temperatures up to 125 K, and
techniques have been suggested to increase this value. Recent theoretical efforts predict
•1 that room-temperature superconducting materials, if possible at all, may require different
chemical compositions than those oxides presently under investigation. A second part of
the report reviews the possible direct use of IITS as infrared sensor elements. This
indicates little or no advantage over semiconductors; however, HTS use in the signal-*1 processing circuitry of detector arrays may be highly beneficial. A second (later) report
will describe a quantitative analysis on the feasibility of a self-guided probe weapon using
superconducting magnetic field gradient sensors. While additional analysis and
calculations are still necessary, the tentative conclusion of this Part I report is that an HTS
seeker and guidance system may be feasible for the destruction of tanks, ships,
submarines, or other large steel targets. Such a missile would be generally insensitive to
the environment or common countermeasures such as flares, camouflage, or smoke
screens.
0
S
4i
IDm
CONTENTS
Preface ................................................................................................. ii
A bstract ............................................................................................... iii
I. INTRODUCTORY REMARKS ........................................................... 1
w II. IDENTIFICATION OF THE HIGH-TEMPERATURE SUPERCONDUCTINGPHASES IN THE VARIOUS MULTIPLE OXIDE SYSTEMS ..................... 3
A. La-Ba-Cu-O System .................................................................... 3B. Y-Ba-Cu-O System .................................................................. 4
C. Bi-Sr-Ca-Cu-O and T1-Ca-Ba-Cu-O Systems ................................... 5
III. CRYSTAL STRUCTURES OF THE HIGH-TEMPERATURESUPERCONDUCTING OXIDE SYSTEMS ............................................ 7
A. Ideal Perovskite Crystal Structure ................................................. 7B. Crystal Structures of the HTS Oxides ............................................ 7
IV. BOLOMETRIC INFRARED DETECTORS ........................................ 13A. Background Discussion ........................................................... 13
B. Superconducting Bolometers ...................................................... 19
C. Conclusions ......................................................................... 22
V. DIRECT SUPERCONDUCTOR INFRARED DETECTORS .................... 23
A. Background Discussion ........................................................... 23
B. Experimental Results (Low-Temperature Superconductors) .................... 25
* C. Experimental Results (High-Temperature Superconductors) .................... 26
D. Conclusions ......................................................................... 30
VI. A NOTE ON SUPERCONDUCTING ELECTRONICS ............................. 31
VII. REFERENCES ............................................................................ 32
iv
I
6
I. INTRODUCTORY REMARKS
• In an earlier IDA report (Ref. 1), a general technical background of the
superconductivity phenomenon was presented, along with a brief history of the
experimental discoveries of increasingly higher critical temperature (Tc) materials since the
publication of the initial data on certain mixed oxides in mid-1986 (Ref. 2). Materials of
* this class have been prepared, to date, with a Tc of 125 K, and there is no reason to believe
this value will not continue to increase, nor is there any reason to believe that such oxides
are necessarily the only type of compounds to exhibit this behavior. Thus, there is an
enormous concentration of research and engineering interest in all technically advanced
* countries in what has come to be called high-temperature superconductors or HTS (as
opposed to conventional low-temperature superconductors, or LTS, which are currently in
use at operating temperatures of about 10 K or less). A condensed version of what is
known about the phases and crystal structures of the HTS materials is presented in
• Chapters II and III.
One of the highly publicized near-term applications of superconductivity is
improved infrared detector systems, a subject on which IDA has a great deal of expertise.
A preliminary analysis is given in Chapters IV and V. An DA concept for using HTS
magnetic field sensors as a detection and guidance system in an anti-tank (or anti-ship)
missile showed promise (Ref. 1). Additional computations, to be published in a
subsequent report, tend to support this promise.
Known HTS materials are multicomponent oxides (i.e., ceramics) having
complicated stability relationships with other compositionally similar, but usually
nonsuperconducting, phases. Knowledge of the phase relations, especially involving
temperature and oxygen pressure, is essential to gaining insight into optimum processing
techniques for these materials. In addition to chemical compositions, the atomic
arrangements (or crystal structures) tend to differ from ideal crystallographic forms; this
difference (and the understanding of it) will provide the basis for future research intended
to create more desirable properties.
* As of this writing, four major superconducting oxide systems have been discovered
(others are known but their critical temperatures are too low to be of interest).
Ii
0
In chronological order, the first HTS compound is generally acknowledged to be
the La-Ba-Cu-O system (Tc about 40 K) discovered by Bednorz and Miller (Ref. 2). The
second is the Y-Ba-Cu-O system of which the most studied is YBa2Cu30 6+x, commonly
referred to as the " 1-2-3" compound, which can have (depending on the value of "x") a Tcas high as 95 K (Ref. 3). The third and fourth compounds are the Bi-Sr-Ca-Cu-O and Ti-
Ba-Ca-Cu-O systems. The latter systems are found to exhibit a T as high as 125 K. The
phases of these four material systems are discussed in Chapters II and III, which also
contain comments about possible future materials.
Considering infrared (IR) detectors, the interest in superconductor detector array
* elements lies in the thought that such devices might offer an attractive alternative to doped
semiconductors for many applications. In particular, space-based systems, when there are
cold targets with a low background, may have to operate in a spectral band up to20 pim or greater. Since this region presents difficulties for the popular mercury-cadmium-
* telluride detectors, extrinsic silicon is usually considered the semiconductor of choice.
Unfortunately, doped Si requires an ambient temperature of around 10 K and thus a
relatively heavy and power-consuming cryogenic system is needed. This is undesirable for
use on board a satellite. If, on the other hand, HTS detectors could operate in the
a temperature range of 20 K to 45 K, the system benefits would be large. It As this type of
consideration which provides much of the motivation (especially in the Strategic Defense
Initiative program) for developing HTS detectors. In the past, work on LTS detectors has
focused on two types of devices: bolometers and so-called direct detectors based on
• Josephson junction tunneling effects. While work on HTS detectors continues to involve
the same types of devices, it must be remembered that the HTS materials are relatively new
and have many unusual properties. It is therefore possible that entirely different IR
detection mechanisms might be developed in the future. This possibility will not be
S considered in this report, which is concerned with current or near-term devices.
Since a major benefit could accrue to focal plane IR detector arrays (as well as an
enormous number of other systems) by the development and use of complex
superconducting electronic integrated circuit devices, a short note (Chapter VI) on this0 subject is included.
2
S
II. IDENTIFICATION OF THE HIGH-TEMPERATURESUPERCONDUCTING PHASES IN THE VARIOUS
MULTIPLE OXIDE SYSTEMS
A. La-Ba-Cu-O SYSTEM
* Although the trend of research interest to the Y-Ba-Cu-O HTS material after its
discovery has led to the relative neglect of the La-Ba-Cu-O and La-Sr-Cu-O oxide systems,
it is important to understand the phases of the latter compounds, if only for comparison.
Among the limited research known to the authors on the phase relations in either of these
* systems is a recently published communication (Ref. 4) of the subsolidus (non-liquid)
ternary phase diagram for the La203-BaO-CuO system at 950'C in air (Fig. 1). The
numbers refer to ratios between LaO1.5, BaO, and CuO.
In the diagram below, the existence of two different HTS compositions can be seen;
one is the material originally discovered by Bednorz and Miiller (La2-x-Bax-Cu-O4, Tc -
36 K), while the other, identified as La3--x Ba3+x Cu6014+8, (x-l), is a material with Tc >
90 K and is crystallographically identical to the Y-Ba-Cu-O 1-2-3 material.
CuO140 I1A
0.011 em
41 (3 123) BCU2
7 2XI) It0.210 422)
LaO 1. 24 BaO
Figure 1. Subsolidus Phase Diagram for the La-Ba-Cu-O System in Air at 9500C.From Ref. 4.
3
0
The subsolidus nature of the diagram provides insight into the possible phases that* can be encountered during solid state processing of ceramics in this system. Since most of
the early fabrication work was accomplished by solid state sintering methods, phase
diagrams at isothermal firing temperatures can be used as guides for the mixing of
constituent oxides. At this time, apparently, no more comprehensive (non-isothermal)
0 phase diagrams have been published for the La-containing HTS materials.
B. Y-Ba-Cu-O SYSTEM
Several investigators have proposed phase diagrams for this widely studied system;
* these are, once again, mostly of the ternary isothermal subsolidus type. Even before the
phase relationships were systematically determined, however, the superconducting phase in
this system was identified as the 1-2-3 compound by Hazen and co-workers (Ref. 5) incollaboration with C.W. Chu of the University of Houston. Hazen also identified a major
* accompanying phase in Chu's samples as a nonsuperconducting 2-1-1 (Y:Ba:Cu)
compound.
Figure 2 is a ternary isothermal phase diagram for the YO1 .5-BaO-CuO system as
determined at 9500C in air by Waig, Mason, et al. at Northwestern University (Ref. 6).
CuO
123: YBa 2OQuO 7 -S
132:YIa3O'hO,, 4
211=Y 2BaCaOs
(123) 0
• "N
YO,.s BaY 20" Ba2Y205 Ba4YzO, aO
Figure 2. Subsolidus Phase Diagram for the Y-BaO-Cu System in Air at 9500 C.
From Ref. 6.
4
Although it differs in some details from the similar diagrams proposed by Frase and Clarke
(Ref. 7) and Roth, et al. (Ref. 8), all of these researchers concur that the HTS 1-2-3
material is essentially a "point" compound, compatible with the 2-1-1 material as well as
CuO and BaCuO2, bvt with no appreciable solid solution range. The stability of 1-2-3
compound, its crystal structures, exact composition, and properties are also highly
sensitive functions of processing, but these aspects will be discussed later. It should also
be noted that the phase equilibria results cited above were preliminary and thought likely to
be revised as further research is done.
C. Bi-Sr-Ca-Cu-O AND TI-Ca-Ba-Cu-O SYSTEMS
To date, several superconducting phases in these newer HTS systems have
evidentally been identified, but no detailed phase equilibrium diagrams published. While
their higher Tc values (up to 120 K), compared to the superconducting La and Y materials,
* sparked great attentions, a number of structural and chemical characteristics make these
newer HTS oxides distinct from the earlier ceramics. Among these are the crystal structure
and the lack of rare earth elements in the Bi and T1 systems. The crystallographic
differences also contribute to a different crystal morphology in the higher Tc materials; for
* example, the bismuth compound exhibit, platey, highly anisotropic habits with cleavage
behavior that makes it appear espec; ally rica-like (Ref. 9).
Another feature that separates these newly discovered HTS oxides from the older
materials is that these systems define classes of HTS phases, rather than just one or two
• compositions. At least one HTS phase has been identified in the Bi-containing system (a
second one is thought to exist as a modification of the first) (Ref. 10), while three have
been identified so far in the thallium-containing materials, including a "precursor" material
lacking calcium (Ref. 9).
Work on the bismuth compounds began in earnest after the discovery by Michel,
et al. (Ref. 11) of 22 K Te superconductivity in the Bi-Sr-Cu-O system.* Later, both
Maeda with co-workers in Japan and Chu's group independently induced HTS behavior in
the Bi system by adding Ca, and Ca and Al, respectively (Refs. 9 and 12). Later work in
the Ca-only containing material identified one of the HTS phases as approximately
*0 Despite the relatively low Tc, this material excited interest because it does not contain a rare earthelement, the first such material discovered in the present class of high-temperature superconductors.
5
Bi 2CaSr 2Cu 20 9 (Tc - 90 K) (Ref. 10). Other material fabricated in Chu's group show
even higher Tc values near 115 K (Ref. 12).
The first of the thallium materials was announced at nearly the same time, an 81 K
Tc T1-Ba-Cu-O ceramic discovered by Sheng and Hermann. The stoichiometry of the
superconducting phase has apparently been determined to be approximately T12 Ba 2 Cu
O6+x (Ref. 13). Sheng and Hermann later added calcium to their material and observed anincrease in Tc to about 105 K, while independent work by Grant and co-workers at IBM-
Almaden confirmed that one of the other phases isolated by Hazen et al. became
superconductive at about 125 K, the highest Tc HTS material discovered to date (Ref. 14).
All three materials are structurally similar to each other (and to the Bi-containing HTScompounds), with the 105 K superconductor identified as T12CaBa2Cu2O8+x and the125 K HTS oxide as T12Ca2Ba2Cu3O10+x. More discussion of how these materials are
structurally related to each other will be given in the next section.
6
HI. CRYSTAL STRUCTURES OF THE HIGH-TEMPERATURE
SUPERCONDUCTING OXIDE SYSTEMS
S
A. IDEAL PEROVSKITE CRYSTAL STRUCTURE
The HTS oxides possess, to varying degrees, crystal structures containing blocks
that are analogous to the class of materials known as perovskites. An idealized crystal
structure is shown by the perovskite unit cell shown in Fig. 3. Perovskites are chemically
expressed by the formula ABX 3 ; a single A cation is located in the center of each unit cell,
with eight B cations at each comer and 12 X anions at the middle of all the cube edges.
Since each comer B cation is shared by eight other unit cells (when the cell is repeatedly in
all three spatial dimensions) and each edge X anion is shared by four other unit cells, the
formula ABX 3 is derived.
In a classic perovskite material, the unit cell symmetry is cubic, but lattice
* distortions that slightly displace ions from their equilibrium positions and change the
overall crystal symmetry are common (Barium titanate, BaTiO3, commonly used as a
dielectric and piezoelectric material, is a well known perovskite that displays such
behavior). The structures of the high Tc superconductors, however, differ from the more
* conventional perovskites in other significant ways.
B. CRYSTAL STRUCTURES OF THE HTS OXIDES
When describing the crystal structure of even the simplest of the high T" materials,
* its complexity requires envisioning stacks of alternating perovskite unit cells such as seen
in Fig. 4 for a "perfect" 1-2-3 compound. A cell containing a centrally located Y (or A) ion
and Cu (B) ions at the comers is sandwiched between two similar cells with Ba ions
located in the central A sites. Oxygen ions are not shown, but would be located at all edge
* midpoints. Oxygen and copper ions are shared uetween adjacent cells. It is this stacking
of three or more perovskite-like cells that must be used to define a new "unit cell" for the
high Tc oxides, which in this particular case, describes a compound with the formula
YBa2Cu309.
7
Figure 3. Unit Cell for an Idealized Perovskite ABX3 .Cubic Symmetry Is Assumed.
Cu01
0
0!
0
0!
Figure 4. Unit Cell for an Idealized Superconducting Oxide, FormulaYBa 2 Cu3Og. (Oxygen ions are not shown, but would normallyreside on all the edges that make up the three perovakite-like
* subcells, midway between the Cu Ions). From Ref. 5.
8
0'
Since the actual HTS oxide known as the 1-2-3 material was found to have the
formula YBa2Cu3O7 ,, it was apparent that defects in the form of missing oxygens were
present. The actual oxygen stoichiometry is believed to be dependent on the presence ofmixed valency Cu ions, but assuming that the Cu is predominantly in the +2 oxidation
state, then the formula for the 1-2-3 oxide is YBa 2Cu306.5 . Most researchers believe,however, that this formula is an average, representing a mixture of materials with oxygen
stoichiometries of both 6 and 7, with the latter being the superior superconductor. Figure 5(Ref. 15) shows the unit cell for this material as it is presently believed to be. Oxygen ions
are missing from all vertical edges of the central yttrium-centered cube and from two of the
four basal plane edges at both the top and bottom. Since this triple-decked cube is nowconsidered the "unit cell" for the 1-2-3 material, each of the eight missing oxygens is
shared with four other unit cells, thus giving the new stoichiometry of YBa2Cu30 7 . (For
the YBa2Cu 30 6 cubic cell, the remaining oxygens in the basal planes are also missing.)
• 9
0. oo
"..i . b -' ........
. 0 .0 ....
.0 . ........c=11,675 .-- .. .
.. 0. .O = Vacancy.........' ; "...... ... 0 . " 0= Oxygen
S...= Copper
a3.884 A
. :b=3.822
Figure 5. Crystal Structure of the HTS Oxide YBA2Cu3O7. The unit celldimensions of this orthorhombic materil are given, while the missingioxygen sites which give its chemical formuia are outlined. From Ref. 15.
9
0I
The crystal symmetry of the 1-2-3 compound is orthorhombic, as can also be seen
from Fig. 5 (i.e., the dimensions of the unit cell are unequal, but orthogonal). The
structure of the La-Sr-Cu-O superconductor differs slightly, being isostructural with the
perovskite known as potassium nickel fluorite, K2NiF 4 (Ref. 15), which possesses a
tetragonal symmetry.
With the advent of the Bi- and Tl-based HTS compounds, the task of describing
their crystal structures becomes much more complicated. Although the concept of"stacking" perovskite-based subcells of different chemistries still applies, there are otherimportant differences that take these new materials further away from the simple perovskite
* unit cell. A sense of this increased complexity can be seen in Fig. 6, which illustrates theatomic arrangements in the the three thallium-based HTS compounds discovered to date,
along with a conjectured fourth compound with an even higher Tc (the bismuth materials
discovered so far have analogous stoichiometries, and are thus thought to have identical
* crystal structures). Once again, these materials have a three-tiered structure; the outerlayers contain mostly oxygen and thallium, while the inner regions consist of one or more
stackings of perovskite-like subcells.
Figure 6 also illustrates a feature of these newer, more complex HTS oxides thatmay provide a clue to the superconducting mechanism in these materials, especially withregard to ever higher Tc values: the existence of stacks of tvo-dimensional copper-oxygenplanes. In the 2-0-2-1 (TI:Ca:Ba:Cu ratio) material containing no calcium, there is only oneCu-O plane, while the 2-1-2-2 and 2-2-2-3 materials have two and three stacked planes,
•respectively. In each case, the increased number of Cu-O planes occurs with a warmer Tc
value, leading to speculation that the presence of more of these planes is a key to raising Tc
even further (hence the extrapolation in Fig. 6 to a fourth 2-3-2-4 compound with Tcestimated at 150 K). In addition, these newer Bi and TI ITS materials lack a feature
0) present in the yttrium 1-2-3 compound that was previously thought to be essential to
superconductivity in these oxides; linear copper-oxygen chains (not illustrated in Figs. 4 or
5). Even the presence of copper-oxygen planes does not appear to be a prerequisite for
superconductivity, as a new non-copper-containing oxide (a perovskite in the Ba-K-Bi-O0 system) with a Tc near 30 K has also recently been discovered (Ref. 16). A recent paper
(W.A. Goddard at an American Chemical Society meeting on September 26, 1988)
predicted that any copper oxide type of superconductor is limited to a maximum Tc of about
225 K, although also stating that the replacement of oxygen by other anions might raise thistemperature.
10
COPPER - OXYGEN PLANES
cv 1 0 01~ cO
0 0
U),rrC.C
0c-0Ln~
cy*caC 3: C
E wZ0 a
6wNr
C0
C E,
=0 0
S0-
so 000 =cy CL0o0
00
ww
- I
all = I
0) 0 0
11E00
These findings in connection with the discovery of new, higher Tc materials appear
to conf'nim the assessment that the mechanism for superconductivity in oxides has not beendetermined. It seems clear that future research directed toward the structural causes of highTc superconductivity may require looking beyond the mechanisms which physicists have
proposed so far (Ref. 17).
40
0
12
IV. BOLOMETRIC INFRARED DETECTORS
A. BACKGROUND DISCUSSION
Bolometers are widely used broadband infrared detectors for slowly varying signals
at far infrared wavelengths. The bolometric effect involves a change in the electrical
resistance of a material due to temperature changes caused by absorbed radiation. As is
well known (see, for example, Ref. 18), the increase in the bolometer's temperature is
governed by the following differential equation
dAT + dPK AT = d A ) (1)
where
C is the heat capacity of the material (JK- 1)
K is the thermal conductance between the element and a heat reservoir (WK- 1)
P is the thermal power generated by 12 Rd heating (W)
A0 is an incremental increase in the power of the incident radiation (W)
e is the detector emissivity (-).
The rate of change of Joule heating with temperature. dP/dT, depends upon the
details of the biasing circuit. However, by assuming an effective thermal conductance
given by• Ke =-K-_d dT
equation (1) can be written as
dAT Ke AT = eA . (2)40 dt
For a sinusoidally varying radiation field
A0 = A00eiWt
13
9
The solution of equation (2) is
* AT = AT exp (- t)K( i) ' (3)
where T (= C/Ke) is the thermal time constant of the device. For typical bolometers, r lies
in the range of milliseconds to seconds. Since the first term in equation (3) represents a
* transient which (under normal operating conditions) can be neglected with no loss of
generality, the change in temperature of the bolometer due to the incident radiation can be
written as
AT = (4)Ke (1 + O2,2) 1/2
From equation (4) it follows that in order to maximize AT, Ke must be small and co << 1.
Thus, the detector's thermal coupling to its surroundings as well as its heat capacity should
* be as small as possible.
Having determined the increase in temperature produced by the incident radiation, it
is necessary to calculate the resulting change in the bolometer's resistance. This in turn
determines the detector's output and so its responsivity. The fractional change in resistance
* due to a change in temperature is given by
ARd 1 -RAT oAT , (5)Rd RddT
where a is the temperature coefficient of resistance. For a simple biasing configuration
(Fig. 7), the output signal is
AV= RL V Rd (6)(Rd+ RL) B
where RL is the load resistance and VB is the bias voltage. Since the detector's
responsivity is defined as
R AVA 0
it follows from equations (4)-(6) that
R =(R ) K IB)RdR LRRRd Ke (1 + 2,C2)1/2 (7)
14
0
where 1B is the bias current. From this equation it can be seen that in addition tominimizing Ke and C, it is desirable to maximize Rd, IB, and a. Unfortunately, these
cannot be varied independently. For example, since Rd must be matched to the input
RL
Radiation - VBOtu01 R
d
0€
Figure 7. A Simple Biasing Configuration for a Bolometer.
impedance of the amplifier, its value is somewhat limited. Moreover, as Rd increases, both
the Johnson noise and the RC time constant (due to the stray capacitance of the connecting
leads) of the device will increase. Likewise, increasing the bias current increases the
detector's 1/f noise. Finally, the temperature coefficient of resistance is, to a large extent,determined by the choice of detector material. Thus, although Rd, IB, and a must be
optimized, minimizing Ke and C provides the most direct way to improve device
performance.
In order to determine the detector's sensitivity, it is necessary to consider the
various noise sources that affect its performance. These can be divided into three
categories: background (or radiation) noise, noise within the detector itself, and noise
associated with the electronics following the detector. Begin by considering the
background noise.
15
2
If the incident background radiation has a mean square fluctuation AW2 , this will
* cause a corresponding voltage fluctuation given byV 2 R2AW 2 (8)
BN
It can be shown that
0 41r(kBTB) 5 A x4e dx
c2h 3 d"f (ex -1)
(Ref. 19), where
TB is the temperature of the background (K)
a is the solid angle subtended by an element of background at the
detector (sr)
* 0 is the angle between the direction of sight and the normal to the
detector (rad)
Ad is the area of the detector (cm2 )
Af is the electrical bandwidth of the system (Hz)
x = hv/kBTB.
The amount of background radiation present is determined by both the angularacceptance of the optical system and the bandpass of any filters used. The noise voltage asa function of background temperature for a 0.2 sr angular acceptance is shown in Fig. 8.Figure 9 shows the noise voltage as a function of the angular acceptance for a backgroundtemperature of 300 K. Since the background temperature is fixed by the scenario in whichthe detector must operate, the only means left to the designer for controlling the level ofbackground noise is the angular acceptance of the optical system. These decisions must bemade on the basis of system level considerations.
1
16
10-
.Sa.
€ I o
C MS 1
0.1.
50 100 150 200 250 300T (K)
Figure 8. Background Noise Voltage as a Function of Background Temperature.
100
z 10
1 IME
0 1 2 3 4 5 6Alpha (sr)
Figure 9. Background Noise Voltage as a Function of Angular Acceptance for aFixed Background Temperature TB = 300 K.
* The detector noise involves three components: Johnson noise, thermal noise, and
1/f noise. Johnson noise results from the random motion of charge carriers in a resistive
element. The noise voltage is given by
* = 4kaTdR , (9)
17
where Td is the temperature of the detector and Rd is its resistance. Thermal noise results
from fluctuations in the temperature of the device, caused by either radiative or conductive
exchange with the surroundings. These fluctuations produce fluctuations in the signal
voltage given by
V = 4kBT'K , (10)
where K is the thermal conductance. The Johnson and thermal noise are plotted as a
function of temperature in Fig. 10. Typical operating parameters for semiconductor
bolometer have been used for these calculations, and are listed in Table 1. As can be seenin Fig. 10, while Johnson noise dominates at low temperatures, as the temperature is
increased the detector's performance is eventually determined by thermal noise in the
device. From equations (9) and (10) it can be seen that the temperature at which thermal
noise dominates is proportional to the detector's resistance (Rd). So as Rd goes to zero (as
30
20*>
C
--- VJN(Td)
10> VTH (Td)
0
0 100 200 300
Td(K)
Figure 10. Johnson and Thermal Noise as a Function ofDetector Temperature
18
0
Table 1. Typical Operating Parameters for aSemiconductor Bolometer
Area of Detector (Ad) 0.15 cm2
Detector Resistance (Rd) 104 ohms
Responsivity (R) 103 V/W
Electrical Bandwidth (Df) 1 Hz
Thermal Conductance (K) 1.83 x 10- 4 W/K
would be the case for a superconducting bolometer), this temperature also goes to zero.
Thus, the fundamental limit to the performance of a superconducting bolometer is given by
the thermal noise in the device. The implications of this fact with regard to bolometers
fabricated from HTS materials will be discussed below.
Both thermal and Johnson noise will occur in an ideal detector. In any real device
there will be additional noise sources, the most important of which is known as 1/f noise.
* Although the mechanism for this noise source is not well understood, it is characterized by
a 1/f power spectrum and is observed to increase with increasing bias voltage. Thus, this
component of the total noise can play an important role in the performance of thermal
detectors. However, it can be minimized by the proper design and choice of operating
* characteristics and will not be considered further here.
Finally, there is noise associated with the electronics which follows the detector.
With the development of low-noise field effect transistor (FET) amplifiers, this noise
source is not as significant as it was in the past, and so will be ignored in what follows.
Assuming that the noise sources are statistically independent, it is possible to write
the total noise voltage as
A useful figure of merit for infrared detectors is the Noise Equivalent Power
(NEP), which is defined as that value of signal power required o produce a signal-to-noise
ratio of unity. It can be shown that
19
NP VaNEP -R (12)
0where Vn is the noise voltage [given by equation (11)] and R is the detector's responsivity.
In low background conditions the detector is limited by either Johnson noise or
thermal noise. In such cases, the ultimate sensitivity of the device can be significantly0 improved if it is cooled to a very low temperature (cf. Fig. 10). Moreover, as the operating
temperature is lowered, the material's temperature coefficient of resistance increases and its
specific heat decreases. These also contribute to a higher responsivity.
B. SUPERCONDUCTING BOLOMETERS
To capitalize on the above effects, some of the first cooled bolometers usedsuperconducting elements. To date, three types of superconducting bolometers have been
developed. The transition edge bolometer is operated at a temperature within the material's
* transition region. Due to the abruptness of the transition, the temperature coefficient ofresistance is quite large, and so the small temperature change produced by the absorbed
radiation causes a dramatic change in the resistance of the detector. It is important to note
that extremely good temperature control is required in order to maintain the device at the
0 transition temperature. In addition to this device, two other types of superconductingbolometers have been built. The semiconductor-normal metal superconductor (SNS)
bolometer uses the temperature dependence of the Josephson current in an SNS junction.
Related to this is a third device, known as the superconductor insulator normal (SIN)
* bolometer, which utilizes the temperature dependence of the quasi-particle current through a
superconductor-insulator-normal metal junction. Except for the junctions involved, the
construction of the SNS and SIN devices is similar to that shown in Fig. 11 for the
transition-edge bolometer. This device consists of a thin film deposited onto a sapphire* substrate, which is covered on the opposite side with a thin bismuth film to absorb the
incident radiation. These so-called composite bolometers were designed to achieve smaller
thermal conductances (thereby improving device sensitivity), while maintaining reasonable
thermal time constants.
0
02
02
Radiation
--- Bismuth
4- Sapphire
*'"Superconductor
Figure 11. A Composite Bolometer
* Table 2 shows the performance that has been achieved to date with the various
types of bolometers. As mentioned above, the theoretical limit of the NEP is determined bythe thermal noise in the device, i.e.,
0 (NEP) = 14kBT2K (13)
Note that the measured values of NEP are given in terms of the electrical (as opposed to
optical) power dissipated. Due to the poor coupling between the detector and the radiation
field, the optical NEP is roughly a factor of two higher than the electrical NEP. For0 comparison, a low-temperature, gallium-doped germanium bolometer is also included. As
can be seen, the transition edge and SNS devices have sensitivities comparable to the
semiconductor bolorneter. In fact, both superconducting and non-superconducting detectors
have nearly ideal NEPs. From equation (13) it follows that further advances depend on
5 building devices with small thermal conductances. However, in order to maintain a
reasonable frequency response, such improvements must be accompanied by
2
21
Table 2. Various Bolometer Characteristics
* Theoretical Observed Time Constant TemperatureBolometer Type NEP (W) NEP (W* (s) (K)
Al Transition Edge( 20) 1.3 x 10-15 1.7 x 10-15 0.08 1.3
Pb/CuAVPb SNS(2 1) 1.1 x 10-15 5.0 x 10-15 3.0 1.5
Pb/A120 3 /AI SIN(2 1) 1.1 x 10- 15 1.0 x 10- 15 1.5
Ge:Ga(22) 2.2 x 10-15 3.0 x 10-15 0.003 1.2
* •The observed sensitivities are given in terms of electrical NEPs (cf. text).
lower heat capacities. Some possibilities along these lines include replacing the bismuth
absorber with a thinner chromium film and the sapphire substrate with diamond (diamond
has a specific heat roughly a factor of 10 less than that of sapphire). At some point,
however, the heat capacity of the film will dominate, and so limit these improvements.
It has recently been suggested that bolometers be fabricated from high-temperature
* superconducting materials. It is important to realize, however, that certain trade-offs are
implicit in such a decision. On the one hand, the ultimate sensitivity of such a device
would be less than that of more conventional bolometers. This is due to two effects. First,
the heat capacity of these materials is somewhat larger than that of conventional
* superconductors. At the point where the heat capacity of the film limits the device
performance, it follows that using high temperature superconductors would in fact lessen
the ultimate sensitivity. A more important consideration, however, concerns the
temperatures at which devices based on these new materials would operate. As discussed
• above, the fundamental limit to the performance of a superconducting bolometer is given by
the thermal noise in the device (i.e., NEP - K 1/2 T). Thus, the higher operating
temperatures involved would degrade the performance. For example, if all else is equal,
the ultimate sensitivity of a bolometer operated at 80 K is NEP - 10-14 W. This is an order
* of magnitude worse than a device operated at 5 K. On the other hand, for satellite
applications in which size, weight, and reliability of the detector's cryogenic equipment is
crucial, bolometers based on high-temperature superconducting materials give the system
designer an additional degree of freedom. For some applications, this consideration may
* outweigh the poorer performance. It must be pointed out in this context that there are
22
0
O
numerous efforts underway to develop room-temperature bolometric devices for certain
* high-volume applications requiring only moderate performance.
C. CONCLUSIONS
Although low-temperature superconducting materials can be used to fabricate
* •extremely sensitive bolometers for low background conditions, the situation does not lookas promising for bolometers made from high-temperature superconductors. The problem
in the latter case involves poor performance resulting from increased thermal noise.
Although it is possible that system level considerations (such as the weight and power
* required by the cryogenic system) could outweigh the lower performance, the consensus
seems to be that bolometers fabricated from high-temperature superconductors do not holdmuch promise. It should be noted in this context that there are numerous efforts underway
to develop uncooled (i.e., room temperature) bolometric devices. It would be interesting to
• compare the performance of these detectors with bolometers based on high-temperature
superconductors.
23
0 ln mn ilnmnm U . .
V. DIRECT SUPERCONDUCTORINFRARED DETECTORS
A. BACKGROUND DISCUSSION
In addition to infrared bolometers, recent efforts have focused on Josephson
•0 Junctions as direct IR drtectors. The detection mechanism in this case is based on the non-
equilibrium state of the superconductor caused by the incident radiation. When optical flux
is absorbed by the material, Cooper pairs are broken, thereby exciting excess quasi-
particles. These quasi-particles will depress the superconducting energy gap. This can be
* seen by the following argument. Consider an electron with momentum p and spin up(pT). Without a corresponding electron in the state (-pl) the pair state (pT, -pl) cannot be
occupied. The electron-electron interaction energy is then smaller since the number of
scattering events in which the electrons can participate is smaller. According to the BCS
0 theory, the superconducting energy gap is given by (Ref. 23)
A = 2hDL [exp (l/p(Ef)M) -1]1 ,
where VL is the average phonon frequency, M is the matrix element of the scattering
interaction and p(Ef) is the density of states for electrons at the Fermi energy. Thus, the
decrease in the pair interaction energy (M) causes a decrease in the superconducting energy
gap.
The Josephson current across the junction can be written as a function of the energy
* gap as follows (Ref. 24)
FAm] L- L 2kT) (14)
• where
A(T) is the energy gap (J)
e is the charge on the electron (Coulomb)
0 Rn is the normal state resistance of the junction (Ohm)
24
T is the temperature of the junction (K).
* Assuming that this equation holds in the non-equilibrium state, it can be seen that
decreasing the superconducting gap will in turn change the I-V characteristics of the
junction. Thus, when such a device is current biased, its voltage is sensitive to infrared
radiation.
• It has recently been suggested (Ref. 25) that IR detection in Josephson Junction
devices is due to a thermnl mechanism. This idea deserves a brief mention. The original
work in this area was done in 1971 by L.R. Testardi (Ref. 26). By studying the dc
resistance of superconducting Pb films subjected to pulsed laser light, he concluded that for
* T < Tc the observed superconducting-to-normal transition induced by the radiation could
not be accounted for by sample heating. Instead, he suggested that excess (i.e., non-
equilibrium) quasi-particles created by photon pair-breaking were responsible. For T >_ Tc,
on the other hand, Testardi found that thermal effects do in fact dominate the material's
• optical response. Following this, C.S. Owen and D.J. Scalapino calculated the energy gap
of a BCS superconductor in which an excess quasi-particle density is maintained by a
radiation field (Ref. 27). It was assumed that the quasi-particles are in equilibrium among
themselves in a Fermi-Dirac distribution characterized by some chemical potential (which isnot equal to that of the Cooper pairs) and a temperature equal to the lattice temperature.
They showed that for T < Tc, the energy gap as a function of the excess quasi-particle
density can be written as
A_= 1-2n
where A is the energy gap of the material in its non-equilibrium state, AO is the energy gap
at zero temperature, and n is the excess quasi-particle number in units of 4p(Ef)Ao, i.e.,
• n= N-NT4 p(Ef) AO
where
0 N is the total quasi-particle density (cm -3)
NT is the thermal equilibrium quasi-particle density (cm- 3)
p(Ef) is the electronic density of states at the Fermi level (cm 3 eV- 1).
•0 This model was confirmed in 1972 by Parker and Williams (Ref. 28). These experiments,
which involved laser illuminated tunnel junctions, found both a decrease in the energy gap
25
and an increase in the quasi-particle chemical potential upon illumination. Since this* original work, further research has also confirmed these results. Thus, one must conclude
that for devices operated at temperatures below the critical temperature, the optical response
is due to the creation of excess quasi-particles and not to a thermal mechanism.
B. EXPERIMENTAL RESULTS (Low-Temperature Superconductors)aIn the past few years numerous research groups have actually built detectors which
operate according to this mechanism. Researchers at the Naval Research Laboratory
(Ref. 29) prepared thin superconducting, granular films of NbN/BN for use as• photodetectors for far infrared radiation (X > 100 gim). The grain boundaries in this
material behave like weak-link junctions and so these films can be thought of as randomnetwork of Josephson Junctions. The detectors, which operate at temperatures less than
7 K, have a spectral bandwidth of 50 gtm to 500 jim. The observed sensitivity of the film
* decreases with decreasing wavelengths. In fact, at visible wavelengths, the responsivity of
the device was roughly two orders of magnitude lower than it was for infrared radiation.
This is due to the fact that the optical absorption in the superconducting film decreases with
decreasing wavelengths. In addition to this broad bandwidth, the devices were relatively* fast, having response times less than 1 ns. The sensitivity of the film was measured with a
pulsed laser operating at a wavelength of 385 .m and a power level of 1 W. The
responsivity was estimated to be about 0.2 V/W. The NRL group claimed that since the
thickness of the film used in these experiments was only 7 nm, it may be possible to
* increase the responsivity by increasing the film thickness, thereby increasing the absoluteabsorption in the material. The rms noise voltage was measured (with only ambient
radiation present and the device current biased) with a lock-in amplifier and was found to
be indistinguishable from the noise voltage of the measurement system. This gave an
* upper bound on the noise level of 0.1 ^tV. Based on these measurements, a noise
equivalent power of approximately 5 x 10-7 W was reported. Unfortunately, the NRLmeasurement of the detector noise was too crude to be useful in predicting the ultimate
performance of their device.
* Workers at NITs Electrical Communications Laboratory in Japan recently reported
using a perovskite-type thin film superconductor, BaPbl-x Bix 03, as an optical detector
(Ref. 30). As in the NbN/BN case, the grain boundaries in the material behave as weak-
link devices, and so are responsible for the direct detection of the IR radiation. The
* Japanese began this work by studying the optical properties of the material. In particular
26
they measured the reflectivity and absorption coefficient for wavelengths between roughly
1 and 20 gm. The dominant feature of this data is the plasma frequency, which occurs at
0.8 pLm (BaPbBiO has a carrier concentration of - 1021 cm- 3, which is roughly two orders
of magnitude less than most metals). For X > 0.8 .m, the material exhibits a high
absorption coefficient (- 105 cm - 1) and a low reflection coefficient (- 60 percent) over a* large spectral band. However, if one assumes that the optical properties obey a Drude-like
theory, (in which the absorption coefficient goes as 1/o2 for frequencies greater than or
equal to the plasma frequency) a X = 0.8 pm plasma edge suggests lower absorption for
visible light and hence fairly low performance in this part of the spectrum. The BaPbBiO
device was operated at a temperature of 6 K. Although the experiment was restricted to
wavelengths from 1 gim to 8 im, the 1 meV energy gap of this material suggests that the
spectral bandwidth could be extended past 500 P.m. The responsivity of this detector,which was measured with a diode laser operating at 1.3 pm, was reported as 104 V/W. Adetectivity of D* - 3.0 x 1010 cm Hzl/2/W was estimated on the basis of a theoretical
speculation concerning the noise in the device. Because the noise in these devices is not
well understood, this value is highly suspect. Finally, the frequency response for the
BaPbBiO device was quite high, with response times less than 1 ns.
* Note that care must be taken when comparing these numbers to the NRL results.
Not only was the film thickness roughly a factor of 20 different, but (more importantly) the
two measurements were done in different spectral regions: the NRL results were obtained
for X - 500 prn whereas the Japanese work was performed at , - 5 pm. Needless to say,
* the optical properties (in particular the absorption coefficients) vary widely over such arange. Whether or not such differences can account for the size of this discrepancy is still
an open question.
* C. EXPERIMENTAL RESULTS (High-Temperature Superconductors)
The Japanese results are very encouraging in light of the discovery of high
temperature superconductivity in the compounds La2-xBaxCuO4 and YBa2Cu307-x. Like
BaPbPBiO, these materials have a low electron concentration, low reflectivity, and a high
* absorption coefficient in the infrared and hence may also be sensitive IR detectors.
Numerous groups are trying to develop IR detectors based on high-temperature super-
conducting materials. They include the following organizations:
27
0
Naval Research Laboratory* Los Alamos National Laboratory
Sandia National Laboratory
National Bureau of Standards (Boulder)
Stanford University
0 Cornell UniversityUniversity of AlabamaWestinghouse Corporation
Hypress Corporation
0 TRW Corporation
Honeywell Corporation
Hughes Aircraft Company.
Some of these efforts are briefly summarized below.
Dr. M. Leung and collaborators at the Naval Research Laboratory recently reported
using a thin granular film of YB2Cu307-x as an optical detector (Ref. 31). The film had atransition temperature of 35 K, with an onset temperature of 80 K. They measured the
temperature dependence of the detector's response and found that the curve had a maximumnear 30 K. This is thought to result from the temperature dependence of the Josephson
coupling between grains in the material. This effect is interesting in light of the discussion(referred to above) concerning the possibility that the optical response of these films isdominated by a thermal mechanism. Since a bolometric response is proportional to the
& material's temperature coefficient of resistance (dR/dT) (cf. equation 7), one would expect
the detector's response to peak in the middle of the transition region, which in this case is
roughly 60 K (cf. Fig. 1 of Ref. 31). Instead, the response drops from a maximum value
of 70 j±V at 30 K to 20 g±V at 60 K. This provides further evidence that the detection
mechanism in Josephson Junction IR detectors (operated at temperatures below the
transition temperature) is not thermal in origin. A blackbody source at 900 K was used to
measure the detector's responsivity and noise equivalent power. Unfortunately, they used
quartz windows on the Dewar which turn-off from 3 pm to 60 p.m (Ref. 32). Thus, when
determining the total optical power incident on the device they had to convolve the quartz
transmission spectrum with that of the blackbody source. How carefully this was done and
what effect it may have had on the measurements was not stated. This notwithstanding,
they reported detectivity of D* - 106 cm Hz1 2/W. This poor performance was due to the
high noise voltage that was present in the device. Examination of the noise characteristics
28
showed that the noise voltage was proportional to the square root of the bias current,
thereby indicating that the noise was in fact shot noise. Workers at NRL believe that this
noise arose from non-ohmic electrical contacts to the device. Thus, as has been the case
with all of the work reported to date on the performance of superconducting tunnel junction
IR detectors, there is no clue concerning the intrinsic noise sources in these devices.
Without such data it is impossible to assess the fundamental limits on the performance of
these novel IR detectors.
The Air Force Office of Scientific Research (AFOSR) has recently begun funding
work at Stanford University on superconducting IR detectors. Early in the first phase of
* this work they were able to control the orientation and microstructure of thin films of
YBa2Cu 3O.x, and found that the electrical and optical properties of the different films were
quite diverse. Thus, it was decided to prepare numerous films, each with a well-
characterized microstructure and then to compare their optical properties. Although this
0 work is currently in progress, some results have been obtained. The optical properties of
the so-called "123 compounds" seem to be very similar to those of BaPbBiO. In particular
the reflection spectra can be explained by a simple Drude theory with a plasma edge at
2.6 eV (7, = 0.5 gtm) (Ref. 33). Thus, in the spectral range 2 gm < X < 10 pgm the
0 reflectivity falls slowly from 90 percent (at 10 gtm) to 60 percent (at 2 p.m). As one
approaches the plasma edge, it begins to fall faster and then levels out to roughly 10 percent
in the visible portion of the spectrum. While this sounds promising for potential
applications in the visible, it should be remembered that if the absorption also behaves
* according to a Drude theory, it will fall off as 1/0 2 in the visible. Thus, the device
performance in this part of the spectrum may be considerably worse.
It should be noted that large sample-to-sample variations in the optical properties of
these materials were observed, despite the fact that the films showed very similar transport* properties. There appear to be two main reasons for this variation. First, the as-grown
films often have non-superconducting phases on their surfaces. By removing this impurity
phase, a high reflectance can be obtained. The second source of variability involves the
different orientations of the various grains present in the films. This anisotropy in the0 optical properties is particularly severe for mid-IR radiation--it decreases as the photon
energy is increased and essentially disappears for X < 6000 A.
The Stanford group has also determined that the optical penetration depth for mid-
IR radiation is around 1000 A. For lower frequencies, this depth decreases due to free
carrier absorption. This suggests that for optimum device performance one should use
29
films whose thickness is 1000 A or less. This is in contrast to previous work which
suggested 1 to 10 pm as the optimum thickness.
Another effort funded by AFOSR is being carried out by Westinghouse. This work
will be divided into three phases and will stretch over the next two years. In the initial
phase Westinghouse hopes to develop both a patterning process for high-temperature
* superconducting thin films and low-resistance, low-noise contacts for these materials.
They will also start measuring the optical response of currently available thin films. At
present, 90 percent of this initial phase has been completed. In particular, pattern
linewidths of 5 .un are now routine and (more importantly) the patterning process does not
* degrade the superconducting properties of the thin films. Unfortunately, the optical
response of the devices is somewhat disappointing in that only a bolometric response has
been observed. Westinghouse attributes this to the absence of any weak-link structure inthe films. Preliminary measurements at X = 0.6 pIm gave a responsivity of 2000 V/W and a
• detectivity of 107 cm Hz1 /2 W-1. Because these results are preliminary, Westinghouse is
unable to account for the high noise levels that are dominating device performance. They
are, however, currently refining their noise measurement capabilities.
The second phase of the Westinghouse effort involves investigating the correlations
* between the film's granularity and its optical and superconducting properties. In particular
they hope to be able to "dope" the grain boundaries in the material in order to control the
nature of the intergrain (i.e., Josephson) coupling. This research is currently in progress
and no results have been reported. The final phase of this work will involve comparing the
• performance of IR detectors based on high-temperature superconducting materials with
semiconductor devices.
Los Alamos National Laboratory (LANL) has recently received a grant from SDIO
to study IR detectors based upon high-temperature superconducting materials. This work
will include various materials systems and will involve thin film materials development aswell as detector characterization. To date, LANL has made a 3 x 3 test array in order to
determine how to make low noise contacts to these materials. Although no performance
data has yet been taken with this device, they were able to make low noise contacts to the
test array. This result, coupled with that from Westinghouse, suggests that making contact
to this material may not be as difficult as originally thought. In order to do further testing
low noise pre-amps will be required. For this aspect of the work LANL will be teaming
with Aerospace Corporation. With a low noise front-end they will be able to investigate the
intrinsic sources of noise in these devices. The importance of such measurements cannot
30
be overestimated--without an understanding of the intrinsic noise in these devices it is
0 impossible to determine the fundamental limits on the performance of superconducting IR
detectors.
D. CONCLUSIONS
* The situation for direct superconductor IR detectors is not clear. As is well known,
the ultimate performance of any IR detector occurs when the noise generated by the
fluctuations in the background flux incident on the device is larger than the thermally
generated noise within the detector itself. This suggests that the detector noise, not its
0 responsivity, drives device performance. The responsivity is important only in that it must
be large enough to ensure that the sensor is not system noise limited. This fact seems to
have escaped many investigators currently working in this field. As a result, the only
measurements reported to date involve the detector's responsivity, with only lip service
1 being given to noise measurements. In the case of HTS IR detectors this lack of data is
especially troublesome. Recall that these devices are being sold on the premise that they
will allow higher temperature operation. But the operating temperature for any IR detector
is determined by the amount of thermal noise that can be tolerated for a given application.
P • To date, there has been no satisfactory work reported in this area. Until this question is
resolved, other efforts in this field seem premature. With this in mind, IDA has recently
begun to consider the thermal noise in superconducting IR detectors. While the results are
interesting, they are very preliminary and will be reported later.
•0
31
0
VI. A NOTE ON SUPERCONDUCTING ELECTRONICS
It has been remarked that one of the most important applications of high temperature
superconductivity in infrared detection may be in the area of signal processing. A typical
focal plane array requires the following "backplane electronics":
Photocurrent integration
40 Charge-to-voltage conversion
Multiplexing
A/D conversion
• Gain and offset corrections
Background subtraction and clutter suppression
Filtering, track formation, etc.
SThe first three items are, in general, located on the focal plane and so are operated at
cryogenic temperatures. In order to achieve a total power dissipation of less than 1 W, it is
necessary that each channel of the device dissipate something less than 10 1W. Given their
high speed and low power dissipation, superconducting electronics would be ideally suited0 for such applications. Moreover, since the A/D converters in these systems consume far
too much power to be located on the focal plane, the analog signals must be brought out of
the dewar to the remaining electronics. The capacitance of these leads can invite a variety
of problems. If the A/D converter could be fabricated with superconducting materials, it
could be moved onto the focal plane, thereby greatly simplifying the sensor. It is important
to note that this would be an important contribution to any IR system, regardless of the type
of detector. Unfortunately, fabricating low-noise Josephson Junction devices is difficult,
even with conventional superconductors. Thus, such applications will be realized in high-
temperature superconductors only in the long term. One interesting possibility along these
lines might be to use a semiconductor detector (such as extrinsic silicon), and low-
temperature superconducting materials for the signal processing. This combination would
be achievable in a much more realistic time frame.
32
0
VII. REFERENCES
0 1. B. Balko, L. Cohen, R. Collins, J. Hove, and J. Nicoll, "Central Research ProjectReport on Superconductivity (FY 1987)," IDA Memorandum Report M-468, May1988.
2. J.G. Bednorz and K.A. Muller, "Possible High Tc Superconductivity in theBa-La-Cu-O System," Zeit. Phys. B, 64, 189-93, 1986.
3. M.K. Wu, J.R. Ashburn, C.J. Tomg, P.H. Hor, R.L. Merg, L. Gao, Z.J. Huang,Q. Wang, and C.W. Chu, "Superconductivity at 93 K in a Mixed-Phase Y-Ba-Cu-OCompound System at Ambient Pressure," Phys. Rev. Lett., 58, 908-910, 1987.
4. D.Klibanow, K. Sujata, and T.O. Mason, "Solid Phase Relations at 950 'C in La-Ba-9 Cu-O," J. Am. Ceram. Soc., 71, C-267 - C-269, 1988.
5. Robert M. Hazen, "Perovskites," Sci. Am., 74-81, June 1988.
6. G. Wang, S.J. Hwu, S.N. Song, J.B. Ketterson, L.D. Marks, K.R. Poeppelmeier,and T.O. Mason, "950 °C Subsolidus Phase Diagram for Y203-BaO-CaO System in
* Air," in Adv. Ceram. Mater., 2 [3B, Special Issue], 313-326, 1987.
7. K.G. Frase and D.R. Clarke, "Phase Compatibilities in the System Y203-Ba-O-CuO," in Ref. 6, 295-302.
8. R.S. Roth, K.L. Davis and J.R. Dennis, "Phase Equilibria and Crystal Chemistry in40 the System Ba-Y-Cu-O", in Ref. 6, 303-312.
9. Ron Dagani, "New Class of Superconductors Pushing Temperatures Higher,"C&EN, 24-29, 16 May 1988.
10. R.M. Hazen, et. al., "Superconductivity in the High-To Bi-Ca-Sr-Cu-O System:Phase Identification," Phys. Rev. Lett., 60, 1174-1177, 1988.
11. C. Michel, et.al., Z. Phys. B, 68, 421, 1987.
12. C.W. Chu, et. al., "Superconductivity up to 114 K in the Bi-Al-Ca-Sr-Cu-O Systemwithout Rare Earth Elements," Phys. Rev. Lett., 60, 941-943, 1988.
9 13. R.M. Hazen, et. al., "100-K Superconducting Phases in the TI-Ca-Ba-Cu-OSystem," Phys. Rev. Len., 60, 1657-1660, 1988.
14. P. Ghosh, "Thallium puts Superconductivity on the Road to 200 K," New Sci.,28, 7 April 1988.
33
15. D.R. Clarke, "The Development of High-Tc Ceramic Superconductors: AnIntroduction," in Ref. 6, 273-287.
16. R.J. Cava (AT&T Bell Laboratories), "Crystal Chemistry and Physical Properties ofHigh Tc Superconductors," given at 90th Annual Meeting of the American CeramicSociety, Cincinnati, Ohio, 2 May 1988.
17. T. Cox, "Superconductivity Discovers the Chemical Bond," New Sci., 64-66,23 June 1988.
18. T. Limperis and J. Mudar, The Infrared Handbook, edited by G.J. Zissis and W.L.Wolfe, Office of Naval Research, Department of the Navy, Washington, DC, 1985.
19. E.H. Putley, Infrared Phys., 4, 1, 1964.
20. K. Rose, IEEE Trans. Electron Devices, ED-27, 119, 1980.
21. J. Clarke, G.I. Hoffer, P.L. Richards and N.-H. Yeh, J. Appl. Phys., 48, 12,4865, 1977.
22. J. Clarke, G.I. Hoffer, and P.L. Richards, Rev. Phys. Appl., 9, 69 (1974).
23. C. Kittel, Quantum Theory of Solids, John Wiley & Sons, New York, 155, 1963.
24. V. Ambegaokar and A. Baratoff, Phys. Rev. Lett., 10, 486, 1963.
25. F. Quelle, Small Device TIWG Meeting; W.J. Schafer Co., Arlington, Va., May 24,1988.
26 L.R. Testardi, Phys. Rev., B4, 2189, 1971.
27. C.S. Owen and D.J. Scalapino, Phys. Rev. Lett., 28, 1559, 1972.
28. W.H. Parker and W.D. Williams, Phys. Rev. Lett., 29, 924, 1972.
29. M. Leung, U. Strom, J.C. Culberston, J.H. Claassen, S.A. Wolf and R.W. Simon,Appl. Phys. Lett., 50, 23, 1691, 1987.
30. Y. Enomoto and T. Murakami, J. Appl. Phys., 59, 11, 3807, 1986.
31. M. Leung, P.R. Broussard, J.H. Claassen, M. Osofsky, S.A. Wolf, and U. Strom,Appl. Phys. Lett., 51, 24, 2046, 1987.
32. W.L. Wolfe: The Infrared Handbook, ed. by G.J. Zissis and W.L. Wolfe, Office ofNaval Research, Department of the Navy, Washington, D.C., 1985.
33. I. Bozovic, et.al., Phys. Rev. Lea., 59, 2219, 1987
34